Overcoming the Challenge of Multi-Device Ad Campaigns Using AI
October 20, 2016 | Marketing 2 minutes read
Online shopping habits of consumers has evolved over the years, along with the development of technology and proliferation of mobile devices. As a result, a consumer’s attention is now fragmented across various devices such as smartphones, tablets, computers and TVs.
According to a recent study by Google, 90% of consumers use multiple devices to carry out a single online activity, while 65% of consumers begin online shopping activities on a smartphone, then continue on other devices. For consumers this might mean convenience, but for advertisers this has created a huge challenge, resulting in wasted ad spending.
The use of artificial intelligence to unify multiple devices to a single user through “probabilistic matching” helps advertisers build more complete consumer intelligence, make better advertising decisions based on a more holistic view of a consumer, and have more moments to influence their target consumer — no matter what device they are using. This puts marketing budget to good use, by actually reaching consumers and not basing campaigns on numbers of impressions.
Probabilistic matching is done by identifying a consumer using machine learning and data mining techniques that analyze signals — including frequency of device usage, location, IP address, web content and app usage behavior, and time of day over seven to 14 days. By using this method, and actively changing how attributions are found and measured across devices, conversion rates can increase by 30 to 40%.
Advertisers are increasingly realizing the impact of probabilistic matching on multi-device marketing. Drawbridge, a company specializing in matching users across devices using probabilistic matching technique, says it has linked 1.2 billion users across 3.6 billion devices — up from 1.5 billion devices just a year ago.
Although programs like Blur prevent consumers from being tracked, especially against deterministic tracking, probabilistic matching is hard to stop. This is mainly because it’s the sum of one’s daily online activities gathered through numerous sources by “Big Data” companies. This increases the accuracy and dependability of probabilistic matching. A recent Nielsen survey found data gathered from Drawbridge to be 97.3% accurate in linking two or more devices, and a similar survey by Nielsen on Tapad data found 91.2% accuracy in probabilistic matching.
Though artificial intelligence through probabilistic matching has immensely increased the success of managing ad campaigns across multiple-devices and skirting around issues such as privacy, it is still in its initial stage of adoption. Many advertisers have yet to fully adopt this technique to manage multi-device ad campaigns.
Going forward, as the number of devices multiplies at a rate that far outpaces growth in the number of actual consumers, marketers must realize that they have to market to people, not devices. Probabilistic matching enables advertisers to accurately re-target consumers, increasing their overall experience as well as increasing their own return on ad spend.
